Flip Chip 焊点可靠性分析

Scott F. Popelar
Delco Electronics Corporation

Abstract

A solder fatigue model for 63Sn/Pb solder has been developed by combining nonlinear finite element modelling with thermal fatigue data of assorted flip chip assemblies. The model characterizes the creep fatigue phenomena of the solder alloy by correlating the amount of creep strain energy dissipated per thermal cycle with the characteristic Weibull life of the critical solder joint. It has been validated for various die sizes, bump geometries, board materials and thermal profiles. Furthermore, the model has accurately predicted fatigue lives for flip chip assemblies with and without underfill. The solder fatigue model has been utilized to investigate the reliability of flip chip joints subjected to thermal cycling. In particular, a parametric study had been performed which shows how various flip chip design parameters will affect solder joint fatigue. Finite element models have been developed to analyze the effect of die size, die thickness, solder joint height, cap diameter and underfill properties on solder fatigue. For this investigation, all analyses have been carried out for parts on ceramic substrates. The results for underfilled parts show that while die size does not influence solder joint reliability, the effects of underfill CTE are very important. Non-underfilled parts are significantly influenced by die size, cap size and joint height.

popelar1.pdf (60.87 KB)

这个方面的文章不少，仿真的也很多。老兄认为这方面的研究前景如何？

去年就想写这方面的文章，当时是想参加ANSYS的年会，

只是没有赶得及完成，目前又想准备一下了，打算投到明年热相关的会议上去，

老兄给点指点吧，谢谢先！

我觉得现在thermal cycling 风头已经过去了，现在大家都在做power cycling 还有mechanical shock loading, drop test 什么的

我是这么考虑的，POWER CYCLE是更为接近实际的情况，小珏有什么IDEAR吗？

最好有cross section的实验观察, 最后的裂纹在IMC里还是bulk solder, 还有最先损坏的位置情况。如果光做power cycling 模拟没有实验的话，新颖度比较少。